The NYT's rent vs buy calculations

David Leonhardt's Economix column has finally been promoted from the front of the Business section to the main front page! Congratulations to him. And the subject matter is dear to my own heart: rent vs buy calculations. In fact, by far the best thing about the article is the online rent vs buy calculator – bookmark it, and use it whenever you or your friends are thinking of buying a place. It's great.

Given his space constraints, one can forgive Leonhardt not going into gruesome detail about all the different variables which go into such calculations. But I would still take issue with a large chunk of how his story is framed.

To read the article, the main variable in determining whether or not you should rent or buy is the amount by which property prices are going to rise in future. Most of the calculations hold everything else constant, and then wonder how many years it will take you to break even given different rates of property-price increase.

But spend a bit of time fiddling around with the calculator, and you realize it's not nearly as simple as that. For instance, the NYT's calculations have a default rate of rent increase of just 4% per year. That seems low to me, given the fact that rent increases haven't remotely kept up with price increases in most of the country. If the two come closer into line with each other, some of that might come from prices going down – but a large chunk of it might come from rents going up. It's hard in the rent vs buy calculator to account for the risk that your rent will suddenly go up by 15% next year.

There are lots of other variables you'll probably want to change, too, like your marginal tax rate (the NYT assumes it's only 20%); the upfront costs of renting, in terms of broker's fee and whatnot (NYT assumes zero, and, it seems, also assumes that renters won't move house any more frequently than owners); and the inflation rate – which has a surprisingly large effect on the rent vs buy curve, for reasons I don't fully understand.

There are three main points I'd make which Leonhardt ignores. The first is that he assumes you have your entire down payment sitting around in a brokerage account compounding at 5% per annum for as long as you have your place. I don't think that's entirely realistic – check out my blog on buying as a commitment device for much more along such lines.

The second point is that when you look at the y-axis, your potential downside is pretty small compared to your potential upside. What the rent vs buy calculator can't do is assign various probabilities to various outcomes and then come up with a net expected return. If it could, buying would become more attractive because of the small chance of a big windfall.

Finally, it's worth noting that the maximum downside is normally pretty much equal to the combined buying and selling costs of owning. The NYT assumes that the cost of buying a house is 4% of the purchase price, and that the cost of selling a house is 6% of the purchase price. When you add those two up, they account for essentially all of the advantage that renting has over buying.

In other words, take your eyes off the house-price appreciation at the exclusion of everything else, and definitely ask yourself what might happen if, for example, the internet helps drive selling costs down to 2% from 6%.

Posted by Felix at 8:58 EST